Flow induced excitation on basic shape

structures
S. Franzetti1 , M. Greco2 , S. Malavasi1 & D. Mirauda2
1 Department

2 Department

I.I.A.R., Politecnico di Milano, Italy.

I.F.A., Basilicata University, Italy.

Abstract
The study of flow-induced excitation on structures and obstacles is one of the
main topics of fluid dynamics related to the practical interests in a large number
of engineering applications e.g. aerodynamic, mechanical, civil, naval, etc. New
design and project techniques have offered hazardous solutions, resulting in structures that are even more slender and flexible. This has led to a number of situations
of self-excited vibration due to the interaction between flow fields and structures.
Forces coming from this mechanism depend upon both the incoming flow and
the structure motion, giving rise to a strong non-conservative force field, which
may eventually lead to a growing structure motion. The aim of this chapter is to
offer an overture about the phenomenon of the fluidstructure interaction. Because
of the importance that the cylindrical and spherical shapes have in the practical
applications and the generalizations that these shapes allow, in this chapter the
fluidstructure interaction is mainly referred to these basic shapes.

1 Introduction
Flow-induced excitations of bodies, obstacles and structures in steady or unsteady
flows, are at present both a relevant field of research as well as the subject of important studies of theoretical and experimental nature.
International literature reports several studies and contributions relating to such
topics for the quasi two-dimensional systems and are summarized in the works
of Sarpkaya [1], Ramberg & Griffin [2], Bearman [3], as well as in the papers of
Blevins [4] and Naudascher & Rockwell [5].
From the 1970s up to the 1980s, the research was mainly focused on the
study and analysis of flow fields and vortex structures generated downstream of

2 Vorticity and Turbulence Effects in Fluid Structure Interaction

the bodies. The emphasis of the results was aimed at a clear definition of the kinetic
characteristics of the currents, related also to the different geometries of the flow
field, through a range of values of several dimensionless parameters governing the
process, such as Reynolds number, Strouhals number and Keulegan-Carpenters
number (Bearman [6], Keulegan & Carpenter [7], Sarpkaya [1]).
Subsequently, from the 1980s on, the development of new acquisition and visualization techniques for describing flow field structures, as well as the increase
in computational capacities for data processing, allowed the research to study by
implementing physical experiments the direct assessment of the effects induced by
the flow fields on the bodies (Blackburn & Henderson [8], Lin et al [9], Sheridan
et al [10]) and the dynamic responses of the obstacle. In these studies the body is
thought as a boundary condition for the flow field.
Only in the last few years, the description of the interaction between body
and flow presents a different approach. It is focused on the possibility of explaining
the different behaviors of bodies in water, and in fluids in general, by looking at
the system as a whole. From this point of view, the body, thought of as a structural
system, does not represent only one of the boundary conditions for the flow field
but, due to its geometrical and mechanical characteristics, plays a relevant role in
governing the dynamics of the process as well.
In order to point out the main active phenomena in the flowstructure interaction
processes, the present chapter deals with the analysis of the sources of excitation
acting on the structure, both external and self-excitation, and the dynamic response
of the obstacle.

2 The source of excitation kinematic implications of flow

structure on induced excitation
2.1 Fluidstructure interaction
The immersion of a solid body in a turbulent flow induces distortions that are
connected to strong kinematic and dynamic instability. The fluid dynamic forces,
due to the fluidstructure interaction, can be analyzed in terms of mean and
instantaneous components; the latter is responsible for the excitation of vibrations.
According to the dominant excitation mechanism involved (Naudascher [11])
the sources of such vibrations can be classified into four groups (fig. 1): (EIE) Extraneously Induced Excitation caused by fluctuating velocities or pressures which
are independent of any flow instabilities originating from the structure and from
structural motion, with the exception of added-mass effect; (IIE) Instability-Induced
Excitation caused by an instability of the flow due to the presence of the structure;
(MIE) Movement-Induced Excitation due to fluctuating forces arising from movements of the vibrating structure; (EFO) Excitation due to Fluid Oscillation caused
by a fluid oscillator becoming excited in one of its natural modes. In any of the
first three cases (EIE), (IIE) (MIE), the exciting forces may or may not be affected
by the simultaneous excitation due to fluid oscillation (EFO). However, even if the
sources of excitation are usually studied according to the above classifications, the

Flow Induced Excitation on Basic Shape Structures

Figure 1: Examples of sources of excitation mechanisms (Naudascher [11]).

excitation of flow-induced vibrations in a real system is very often complex, since

EIE, IIE, MIE and EOF may occur simultaneously.
In fig. 1, the Extraneously Induced Excitation (EIE) is represented by the instabilities of the incoming flow due to its turbulent level, which is not affected by the
characteristic of the structure. Other sources of EIE are: earthquakes, machines and
machine parts, two-phase flow and oscillating flow. The Instability-Induced Excitation (IIE) is depicted by the vortex shedding downstream to a stationary circular
cylinder; the shape of the obstacle mainly affects the kinematic characteristics of
the wake. In the case of the Movement-Induced Excitation (MIE), the obstacle is
not stationary; its movements interact with the vortex shedding evolution or induce
vortex shedding. In this situation the structure behaves like a body oscillator (see
Section 3): the transverse movements of the structure induces distortions on the flow
field which in turn induces the self-excitation of the structure. Finally, the Excitation
due to Fluid Oscillation (EFO) mechanism is represented by standing gravity waves
generated between a long pier and the walls of a flume. Flow-induced excitation can
be enhanced by the EFO mechanism especially if one of EFO frequencies assumes
the natural body-oscillator frequency, the dominant frequency of flow instability,
or both (refer to the relevant literature for an extensive discussion on the effects of
fluid oscillations: Guilmineau & Queutey [12], Lam & Dai [13], Yan [14]).
The structure of an external flow around an immersed body and the way in
which the flow can be described and analyzed often depends on the geometry of the
body. Three main categories of bodies are usually considered: (a) two-dimensional
objects (infinitely long and of constant cross-sectional size and shape); (b) axisymmetrical bodies (formed by rotating their cross-sectional shape around the axis of
symmetry), and (c) three-dimensional bodies that may or may not be symmetrical.
In practice there can be no truly two-dimensional bodies, however, many objects
are sufficiently long so that the end effects of considering the body as twodimensional are negligible.
Another classification of body shapes can be made depending on whether the
body is streamlined or blunt. Flow characteristics depend strongly on the amount
of streamlining present. In general, streamlined bodies (e.g. airfoils, racing cars,
etc.) have a little effect on the surrounding fluid in comparison with the effect of
blunt bodies (e.g. buildings, parachutes, etc.).

4 Vorticity and Turbulence Effects in Fluid Structure Interaction

The geometrical shape of structures or vehicles is very complex and the fluiddynamic efficiency of the shape is usually studied employing a physical model.
Even though the fluidstructure interaction mechanism has been extensively
studied on basic shapes, it still presents open questions; therefore a large part of
current studies still concern basic shapes (cylinders, prisms, spheres, etc.), also
because the fluid-dynamic studies on complex structures are hardly extendible to
other shapes or boundary conditions.
The cylinder is one of the basic shapes most studied because of the simplicity
of its form and because this form mimics a large number of practical applications.
Long-span bridges, tall buildings, tall towers, cables, and so on, are examples of flexible cylindrical structures that are very sensitive to vortex-induced vibrations. The
characteristic elongation in the transverse direction makes the cylindrical shapes
very sensitive to the induced excitation of the flow. To be able to find the appropriate countermeasures required to control the fluid-dynamic response, the generation mechanism of this response should be clarified first. This mechanism is
the flow pattern around the obstacle. Because of the complexity of the flow pattern, in order to understand the fluidstructure interaction mechanism, usually basic
cross-sections such as 2-D rectangular cylinders, H-shaped cylinders and circular
cylinders have been investigated and mainly in smooth flows.
In smooth steady flow conditions the cross-sectional dimensions of a stationary
cylinder are generally the main characteristics responsible for the flow pattern
deformations and consequently for the induced excitation on the cylinder. The
sensitivity of the flow pattern on the main parameters that affect the phenomenon
(such as Reynolds number, turbulent intensity, aspect ratio etc.) depends on the
cross-sectional form of cylinder.
While vortex shedding principally depends on the Reynolds number (fig. 2) for
a circular cross-sectional form, the vortex shedding mechanism is more complex for
a sharp edge cross-section. For a rectangular sharp edged cross-section, the aspect
ratio (L/D) generally represents the main parameter to be taken into account.
The forces that the flow induces on a circular cylinder affected by the flow
field structures shown in fig. 2, both on the mean and fluctuating components
are highlighted in figs. 3 and 4. The figures detail the drag coefficient and the
Strouhal number of a circular cylinder versus Reynolds number. The different values of Strouhal number are justified by the characteristics of the boundary layer
on the cylinder and of the near wake. In a subcritical range of Re (150 300 < Re
< 1 105 1.3 105 ), the near wake passing a stationary smooth cylinder is laminar, the vortices shed periodically and the force fluctuations correspond to a
spectrum of extremely narrow bandwidth. In the post critical range of Reynolds
number (1 105 1.3 105 < Re < 3.5 106 ), the boundary layer on the
cylinder becomes turbulent downstream from a laminar separation bubble, and
the near-wake becomes less regular; the force spectra in this post critical range are
rather broadband. In the transcritical range Re > 3.5 106 , finally, the boundary
layer becomes turbulent upstream of separation and there is an apparent return
of well-defined periodicity of both vortex shedding and force fluctuation with
Sh 0.3 (fig. 4c).

Flow Induced Excitation on Basic Shape Structures

Figure 2: Flow patterns for flow past a smooth circular cylinder at various Reynolds
numbers: (a) Re 1.5 101 , no separation; (b) Re 1.5 101 , steady
separation bubble; (c) Re 1.5 102 , oscillating Karman vortex street
wake; (d) Re 2.5 104 , laminar boundary layer with wide turbulent
wake; (e) Re 3.2 105 , turbulent boundary layer with narrow turbulent
wake (Munson et al [15]).

2.2 Sharp-edged rectangular cylinder

In the case of a sharp-edged elongated rectangular cylinder, the flow detaches on the
upstream (primary separation) and downstream corners (secondary separation), and
the flow distortion is affected by several parameters. The vortices generated close to
the body develop and shed from it, creating an unsteady wake. The characteristics of
the wake are dependent on the Reynolds number of the obstacle Re = U0 D/, but
the aspect ratio (L/D) and the incidence angle between cylinder and flow have the
main influence on the vortex shedding and therefore on the structural excitation.
Figure 5 shows the mean flow field around a rectangular cylinder (L/D = 3) in
unbounded flow, numerically obtained by Yu & Kareem [16] at Re = 1 105 .
The primary separation of the shear layer occurs on the leading edge. The separated flow initially diverges from the body, with an angle dependent on
the separation pressure, and then curves toward the cylinder surface. When the
reattachment occurs, the flow makes a region of recirculation known as separation

Figure 3: Drag coefficient as a function of Reynolds number for smooth circular

cylinder and smooth sphere (Munson et al [15]).

6 Vorticity and Turbulence Effects in Fluid Structure Interaction

Figure 4: Strouhal number, Sh, of vortex shedding (a) and spectra of the lift force
component (b, c) from a stationary, smooth circular cylinder in lowturbulence cross flow (Naudascher [11]).

bubble. In such conditions the secondary separation at the downstream edge of the
cylinder causes the roll up of the flow in the rear face generating a secondary vortex
that periodically sheds from the body surface. In some cases, when the reattachment of the primary separation is unsteady, the two turbulent structures interact and
the vortex shedding becomes more complex. To simplify the description of the phenomenon, the main vortex shedding regimes have been defined and classified on
the basis of the characteristics of the main vortices involved. In the case of steady
flow conditions and rigid obstacle, Naudascher & Wang [17] give the following
classification:
LEVS (Leading-EdgeVortex Shedding): the flow separation occurs at the leadingedge with formation of vortices dominating the near wake of the body (fig. 6a);
TEVS (Trailing-EdgeVortex Shedding): a decisive flow separation at the trailingedge occurs and vortex-shedding is analogous to the von Karman street behind
circular cylinders (fig. 6c);
ILEV (Impinging Leading-Edge Vortices): a flow separation at the leading-edge
and impingement of the leading-edge vortices at the side surfaces and/or edges of
the body are present (fig. 6b);

Figure 5: Mean flow field numerically obtained in unbounded flow around a rectangular cylinder, L/D = 3, Re = 1 105 (Yu & Kareem [16]).

Flow Induced Excitation on Basic Shape Structures

Figure 6: Mechanism of vortex-shedding for IIE source condition on prismatic

2D obstacle (Deniz & Staubli [18]).

AEVS (Alternate-EdgeVortex Shedding): both the leading-edge and the trailingedge mechanisms are present (fig. 6d).
Each vortex type allows a specific dynamical state. Under the flow conditions
above mentioned and for a wide range of Reynolds numbers, the aspect ratio (L/D)
affects the vortex shedding and the loading on the structure significantly. When
L/D < 2 (fig. 6a), only the primary separation occurs because the shear layer separates at the leading edge and involves the whole side of the cylinder (LEVS); in
this range of L/D Bearman & Trueman [19] observed that the formation of the
vortex close to the cylinder enlarges the drag coefficient of the obstacle (CD ).
The minimum distance between rear cylinder face and vortex formation occurs for
L/D = 0.64, which corresponds to the maximum value of CD .
When L/D > 6 (fig. 6c), the flow separated at the leading edge reattaches permanently; consequently the trailing edge separation (TEVS) dominates the vortex
shedding. For L/D 2.8 (fig. 6b), the literature indicates a complex situation of possible unstable reattachment (Shimada & Ishiara [20], Yu & Kareem [16]). When
= 0 (fig. 6d), the symmetry of the flow structure is compromised. Consequently,
on one the upper side of the cylinder LEVS prevails and on the lower, TEVS. When
this condition occurs, the vortex shedding is characterized by the AEVS regime.
The vortex shedding behavior is well described by the Strouhal number
(Sh = f0 D/U0 , where f0 is the dominant frequency of the vortex shedding).
The chart in fig. 7 reports the Sh obtained by several Authors with rectangular
cylinders of various L/D immersed in unbounded flows. In fig. 7, two discontinuities
of Sh values are evident. The first occurs when L/D is approximately equal to 2.8,
the second occurs when L/D = 6. At L/D2.8, the flow pattern is bounded between
the flow separation type LEVS and the flow reattachment type ILEV. The data
dispersion in fig. 7 and the presence of more than one dominant frequency for a
specific aspect ratio mainly depend on the upstream flow characteristics.At the latter
critical aspect ratio L/D = 6, the flow pattern is bounded between the unsteady flow
reattachment type ILEV and the completely steady flow reattachment type TEVS.
For a wide range of conditions, several studies show the negligible influence
of Reynolds number when it exceeds the value Re = 1 104 . On the contrary, the
influence of Re is not negligible for Re < 1 104 (Okajima [21]). In fig. 8, the Sh
versus Re is reported with an aspect ratio L/D = 3.

8 Vorticity and Turbulence Effects in Fluid Structure Interaction

Figure 7: Variations of Strouhal number according to L/D ratio for rectangular cylinders in unbounded flow (Shimada & Ishiara [20]).

The free stream turbulent level of the flow passing a circular cylinder,
Tu = u
rms /U0 (where u
rms is the standard deviation of the inflow velocity on
x direction), significantly affects the flow pattern and the excitation induced on
the cylinder. As shown in fig. 9a, turbulence decreases CD at subcritical Re and

Figure 8: Variation of Strouhal number with Reynolds number for rectangular cylinders with L/D = 3 (Okajima [21]).

Flow Induced Excitation on Basic Shape Structures

Figure 9: Effect of the free-stream turbulence, Tu, on (a) mean drag coefficient,
CD , (b) and on Strouhal number, Sh, for a smooth circular cylinder (Naudascher [11]).
increases it in the supercritical range. The rise in value of Strouhal number in the
transition range (fig. 4a) occurs at smaller Re as Tu increases (fig. 9b).
In the case of a sharp-edged rectangular cylinder, free stream turbulence level
(Tu) has received a great attention in literature because it significantly influences
the structure and the development of the shear layer separated off the upstream
corners (Haan et al [22], Lin & Melbourne [23], Noda & Nakayama [24], Saathoff
& Melbourne [25]). The main effect of Tu is to shift the reattachment point. An
increase of Tu leads to a progressive shortening of bubble formation and, thus, to
a possible strong modification of vortex shedding (Nakamura et al [26]). Noda
& Nakayama [24] observed that turbulence shakes the shear layer over a distance comparable with the turbulence scale. The main effects of turbulence occur
when L/D is in a range of values near the critical value L/D = 2.8. In this range
of L/D, the reattachment of the leading edge separation is not stable. In this
situation, the turbulent inflow with the length scale of the same order as D acts
by moving the position of the separated shear layer off the downstream corners,
promoting the reattachment.
The behavior of vortex shedding is significantly affected also by the presence
of boundaries that limit the evolution of the wake. The presence of boundaries are
relevant in a large number of civil applications (e.g. buildings, bridges, pipelines,
etc.)
The study of boundary effects has principally been considered in aerodynamic
applications especially in terms of blockage ratio (b ), defined as the ratio between
the frontal area of the body and the cross-section of the flow without obstacle.
Both for a circular or rectangular cylinder, significant changes in Sh values can
occur when the flow confinements are changed. In general, the increasing of the
blockage induces an acceleration of the flow near the object, which locally increases
the flow velocity (solid blockage) and increases the energy losses in the wake and
in the boundary layer (wake blockage). For bluff bodies, the effects of blockage
(solid and wake) can be very remarkable and its influence changes the values of

10 Vorticity and Turbulence Effects in Fluid Structure Interaction

Figure 10: Strouhal number results versus elevation ratio: Price et al [27],
/D = 0.45, Re = 1900;  Price et al [27], /D = 0.45, Re = 4900;
2 Angrilli et al [28], Re = 2860, 3820, 7640, /D = 0.2, 0.4, 0.5;
Bearman & Zdravkovich [29], Buresti & Lanciotti [30], Taniguchi &
Miyakoshi [31] and Lei et al [32], 1.3 104 Re 1.4 105 ,
0.1 /D 1.64. (Price et al [27]).

both force coefficients and Strouhal number. These effects are usually taken into
account using the following expressions:
CF c = CF (1 b )nCF ;

Shc = Sh(1 b )nsh

(1)

where CFc and Shc are the corrected force coefficient and Strouhal number, respectively and nCF and nsh are experimental coefficients (0 nCF , nsh 1).
The presence of a significant asymmetry of the boundary conditions has also
remarkable effects on the structure excitation. These effects are summarized in fig.
10, where the Strouhal number of a circular cylinder is plotted against the elevation
ratio, G/D, of the cylinder above a wall (G is the elevation above the wall and
D is the diameter of circular cylinder). The effects on the cylinder excitation are
emphasized by low Re values and affected by the boundary layer thickness ()
above the wall.
The influence of a solid surface on the dynamic effects for a rectangular cylinder
has been less considered in literature. A recent study (Cigada et al [33]) highlights
that for a rectangular cylinder with aspect ratio L/D = 3, the presence of a solid surface significantly affects both the force coefficients and the vortex shedding even if
the cylinder is placed at relevant elevation, G, from the surface. The solid boundary
affects the lift coefficient, CL , up to G/D  3.5; in the range 3.5 G/D 1,
CL decreases toward the negative value CL = 1 then increases up to CL = 1 in
the range (1 G/D 0). The influence on drag coefficient seems limited in the
range 1 G/D 0, where CD decreases from its typical unbounded value up to
CD = 0.7.

Flow Induced Excitation on Basic Shape Structures

11

Figure 11: Dependence of the experimental drag coefficient on h and Frs

(G/D = 2.33), together with the (constant) value of CD = 1.3 for
unbounded flow (Malavasi & Guadagnini [34]).

In hydraulic applications the influence of a free surface is usually considered

as one of the main parameters in modelling structure excitation, relevant in the
study and assessment of the vulnerability of river bridges under partial or total
submergence conditions. Experimental studies, carried out by Denson [35] and
Malavasi & Guadagnini [34], highlight the significant influence of the free surface
in terms of force coefficients. The influence of the free surface on the flow field
structure around the obstacle was provided in Malavasi et al [36].
Figure 11 shows the behavior of CD versus h for different value of the Froude
number, Frs , for a rectangular cylinder with L/D = 3 placed at an elevation
G/D = 2.33 from the bottom of a hydraulic channel, where CD is computed as:
CD =

FD
,
0.5U02 D

(2)

h is the dimensionless parameter of the cylinder submersion:

h =

hG
,
D

(3)

Figure 12: Lift coefficient versus h with Frs = 0.26, together with the reference value of CL = 0 corresponding to the unbounded flow condition
(Malavasi et al [36]).

12 Vorticity and Turbulence Effects in Fluid Structure Interaction

Figure 13: Strouhal number, Sh, versus h (G/D = 2.33).

and Frs is the Froude number referred to the cylinder thickness:
U0
F rs = .
gD

(4)

As h increases toward 1, CD increases independently from Frs ; with further

increases of h , CD reaches its peak value then decreases and tends to an asymptotic
value. The peak of CD seems to depend on the observed value of Froude number
and h .
The free surface also drastically influences the lift coefficient, as shown in fig. 12
where CL is plotted versus h (Frs = 0.26). The lift coefficient presents a negative
peak for h = 1, after which the absolute value of CL increases tending to zero
(unbounded flow condition) as h increases.
In fig. 13, the vortex shedding frequency in terms of Strouhal number and calculated by the frequency analysis of the lift component on the rectangular cylinder
is plotted versus h for different Frs . The significant difference between the ex-

Figure 14: Mean flow field reconstruction by velocity field measured around the
cylinder with Frs = 0.26, G/D = 2.33 and h = 1.0 (a) and h = 1.4 (b)
(Malavasi et al [36]).

Flow Induced Excitation on Basic Shape Structures

13

perimental values and the reference value of the unbounded condition may be
explained in the features of the confinement of the flow. As shown for example in
figs. 14a and 14b, the upper confinement of the flow interferes with the leadingedge separation, changing the structure and the characteristic of the vortex shedding.
The asymmetry imposed by the free surface limits the separation on the topside of
the cylinder, thus the lack of equilibrium on the vertical loading direction induces
significant variation in the CL value from CL = 9 to CL = 2 as shown
in fig. 12.

3 Dynamic response of the structure

3.1 Basic equations
The analysis of the interaction between flow and structure may also be put forth
using the behavior of the body as a reference point for characterizing the processes.
This allows us to evaluate the dynamic response of the oscillating obstacle, compared to vortex-induced vibrations phenomenon and the main characteristics of the
flow field.
The equation of motion generally used to represent the vortex-induced vibrations
of a body oscillator, in steady and unsteady flows, is proposed as follows:
m
x + B x + Cx = F (t),

(5)

where x is the displacement of the body towards the main flow or in a transversal
direction (fig. 15a), m is the total structural oscillating mass, B is the structural
damping, C is the spring constant and F is the acting fluid force. In this case,
the body oscillator is treated as a discrete-mass system free to vibrate in one/two
directions and the fluid force assumes a sinusoidal form:
F = F0 sin(s t + s ),

(6)

where s = 2fs is the circular frequency and fs is the frequency of fluid force.
The solution of eqn. (5) is composed by the solution of the homogenous equation
given by:


x = en t x0 cos(d t ) with d = n 1 2 ,
(7)
and the particular solution:
x = x0 cos(s t ),

(8)

where n = 2fn is the natural circular frequency and fn is the natural frequency
of the system.
The frequency fn , can be generally calculated under the hypothesis of perfect
joint taking into consideration the contribution of the added mass according to the
following relationship:

14 Vorticity and Turbulence Effects in Fluid Structure Interaction

Figure 15: (a) Simple body oscillator with linear damping (i.e. resistance proportional to velocity); (b, c) histograms of responses for an underdamped
( < 1) and an overdamped ( 1) case (Naudascher & Rockwell [5]).

fn =

1
2

C + C

,
m + ma

(9)

where C
represents the added stiffness, which is usually included in the spring
constant of the system and ma the related added mass.
The damping factor or damping ratio, , is defined as:
B
B
=

.
2mn
2 (m + ma )C

(10)

For the underdamped case ( < 1) the damping factor can be calculated by the
exponentially decaying response for the initial condition t = 0, x = x0 and = 0
(fig. 15b), as mentioned by Naudascher & Rockwell [5]. In the case 0 < < 1 the
coefficient has been obtained through the following equation:
= ln

2
xn
=

,
xn+1
1 2

(11)

where is called the logarithmic decrement.

For 1, the displaced body simply returns to its equilibrium position in an
exponential fashion (fig. 15c). The damping for the limit case of = 1 is called
critical damping.
Since the solution eqn. (7) dies out with time on account of damping, only the
steady-state solution eqn. (8) is of general interest. Its frequency is equal to the
forcing frequency fs = s /2 and the amplitude x0 is obtained as:
x0 =

F0 /C
[1 (s /n )2 ]2 + (2s /n )2

(12)

The phase angle by which the response x lags the exciting force F is as
follows:
tan =
where n =

2s /n
,
1 (s /n )2

C/m + ma is the natural frequency of the undamped system.

(13)

Flow Induced Excitation on Basic Shape Structures

15

For a body with one torsional degree of freedom, eqn. (5) takes the form:
I + B + C = M (t)

(14)

where I is the mass moment of inertia of the body, M (t) is the exciting moment or
torque, B is the damping moment,
C is the restoring moment, = B /2I n


is the damping ratio and n = C /I is the undamped circular natural frequency.
The response to a harmonic exciting moment M (t) = M0 cos s t is:
= 0 cos(s t )

(15)

where 0 and are the amplitude of torsional vibration and the phase angle respectively.
3.2 Dynamic response in resonance conditions
In the study of flow-induced vibrations a condition of particular interest is the
resonance phenomenon when the vortex shedding frequency is close to the natural frequency of the structure. This occurs because the resonance phenomenon
generates critical conditions in the structures in terms of stability and structural
stress corresponding to potential collapsing of the structures themselves.
Another case of relevant importance is one in which the frequency of the body
oscillations matches the frequency of the wake vortex. In such cases the processes
are outlined as lock-in or synchronization. The body tends to pulse presenting
large amplitude and the system, even if it does not assume the resonance condition, is subject to relevant stress. In such conditions the oscillation amplitudes, transversal to the fluid flow (y-direction), are always found to be much larger than
streamwise motions (x-direction).
Studies on the analysis of the vibrating structures nearing the conditions of
resonance in bounded and free surface flows, have highlighted the existence of
a strong dependence of the maximum transverse amplitude Amax on some nondimensional groups, as shown below:


U0
h

; SG ;
(16)
Amax = A
fn D
D
where Amax = y0 /D is the ratio between the maximum transverse amplitude and
the characteristic dimension D of the body, the ratio fUn0D is the reduced velocity U , h/D is the ratio between the water depth and the characteristic size of the
body and SG is the Skop-Griffin parameter defined as follows:
SG = 2 3 Sh2 (m ),

(17)

where Sh is the Strouhal number fUD0 .

Concerning the influence of the body shape, the SG parameter takes into account
both boundary conditions of the oscillating body, through m (ratio between the
structural mass m and the added mass ma ) and , as well as the flow induced force

16 Vorticity and Turbulence Effects in Fluid Structure Interaction

Figure 16: Griffin plot showing maximum amplitude observed in different experiments versus the combined mass-damping parameter (Khalak &
Williamson [41]).

through Sh. Under resonance conditions, the Strouhal number is assumed constant, thus the maximum transverse amplitude, Amax , depends on the Skop-Griffin
parameter and, as seen in eqn. (17), on the combined mass-damping parameter
m .
Figure 16 summarizes the results of several experiments for different values
of m in terms of maximum transverse amplitude, Amax , versus m . From this
figure, it does not seem possible to make a singular curve of Amax versus m .
Sarpkaya [1] originally stated that a simple observation of the motion equation
immediately shows that the response of the system is independently governed by
mass and damping. By analyzing three pairs of low-amplitude response data, each
pair of them at similar values of m but different m values, he observed a large
influence of mass ratio on Amax . In fact Sarpkaya [43] states: one should use the
combined parameter m only for m > 0.40 while for m < 0.40 the dynamic
response of system is governed by m and independently.
Khalak & Williamson [41] carried on a set of experiments over a wide range
of m (m = 1 20) under the same experimental conditions showing that even
for low m of the order 2 and very low mass-damping down to the value m
0.006, the use of a single combined mass-damping parameter collapses peak amplitude data very well, even for a wide independent variation of parameters m and
(fig. 16). In this way they extended the value of m proposed by Sarpkaya by two
orders of magnitude.
Furthermore, in the case of elastically mounted systems, they observed two
different types of response depending on the high or low combined mass-damping
parameter m . In fact for low m values, there are three different branches of
response: the initial, the upper and the lower ones which present two jumps in

Flow Induced Excitation on Basic Shape Structures

17

Figure 17: Maximum amplitude versus reduced velocity for different bodies:
Khalak & Williamson [41] and Feng [37] on the cylinders; Jauvtis
et al [44] and Mirauda & Greco [45, 46] on the spheres.
the magnitude of oscillating displacement (fig. 17). They found that the transition
between the initial and upper branch was hysteretic, while the transition from
the upper to lower branch involved an intermittent switching.
On the contrary, for high values of combined mass-damping parameter m ,
Feng [37] observed only two branches of response: the initial branch and the lower
one. The passage between the two branches, as can be seen in fig. 17, occurs with
a jump and the body reaches conditions of resonance.
Furthermore, Govardhan & Williamson [42], by visualization techniques
(Digital Particle Image Velocimetry), showed that the change from the initial branch
to the upper one, depends on the jump in the angle phase between the force induced
by the shedding of the main vortex and the displacement of the body (fig. 18). This
jump is characterized by a change in the form of the vortex wake downstream of
the body by a mode 2S, indicating 2 single vortices shed per cycle, to mode 2P,
meaning 2 pairs of vortices per cycle (fig. 18). Under this condition the value of
the body oscillating frequency, f , passes across the natural frequency in water generating a resonance phenomenon. On the other hand, the passage from the upper
branch to the lower one is characterized by the presence of a phase-difference
between the total fluid force and the displacement of the body which tends to go
toward a periodic uniform trend. In such cases no change in the form of the wake
is observed.
For high values of m the passage from the initial branch to the lower one
depends on the jump of a phase between both the force components, the total force
and the force induced by the vortex and such jump is related to a change in the form
of the wake.
Referring to fig. 17, the behavior found for three-dimensional structures, with
elementary geometrical forms (ex. spheres), is sensitively different from that
observed for two-dimensional structures. In fact, the data of Jauvtis et al [44] relating to the oscillations of a sphere, indicate the presence of two distinct modes of

18 Vorticity and Turbulence Effects in Fluid Structure Interaction

Figure 18: Example of flow visualization of the initial branch (2S mode) and the
lower branch (2P mode) (Khalak & Williamson [41]).

response. The first mode of response (Mode I) is manifested in the presence of resonance conditions, when the frequency of the shedding of the vortex is close to the
natural frequency of the body, and a synchronization regime is observed between
the force and the response. When the average velocity of the flow increases, the
system shows the presence of periodic oscillations characterized by high values of
displacement that represent the second mode of response (Mode II).

Flow Induced Excitation on Basic Shape Structures

19

Figure 19: Frequency ratio versus reduced velocity for vibrating cylinder.

In fig. 17, data from Mirauda & Greco [45, 46] are also reported. The first set
(squares), referring to a steel sphere in a free surface flow with a high value in
the combined mass-damping parameter, is characterized by low oscillations and
show only the initial branch without a jump in amplitude and, therefore, they do
not exhibit hysteresis phenomena. The second series (triangles), characterized by
values of m lower than the previous ones, are close to the first mode of response.
It outlines how the system tends to reach the resonance conditions where vortexshedding frequency is equal to the natural frequency.
The results reported in fig. 17 can be better outlined by referring to figs. 19
and 20 which report the values of the f , ratio between the body oscillating frequency f and the body natural frequency fn , versus the reduced velocity U . In
particular fig. 19 shows the data observed by Khalak & Williamson [41] for vibrating circular cylinders with mass ratio equal to 2.4, 10.3, and 20.6 and fig. 20 the
data observed by Jauvtis et al [44] and Mirauda & Greco [45, 46] for vibrating
spheres with a mass ratio equal to 80, 7.9 and 1.14, respectively.
In the figures, the horizontal line represents the condition in which the oscillating
frequency f is equal to the natural frquency and the diagonal line is the condition
in which f is equal to the vortex-shedding frequency for the static cylinder.
It has been observed that for low mass ratios, oscillation frequency starts from
the natural frequency as the velocity U increases and this transition is characterized
by the presence of hysteresis.
On the contrary, in the case of high mass ratios the synchronization regime
decreases and the value of f remains close to the unity for all values of U .
This is true both for the two-dimensional structures (cylinders) as well as for
three-dimensional structures (spheres).

20 Vorticity and Turbulence Effects in Fluid Structure Interaction

Figure 20: Frequency ratio versus reduced velocity for vibrating spheres.
In the case of free surface flow, the dynamic response is also conditioned by
the parameter h/D. In fact, fig. 21 shows the experimental results of Mirauda &
Greco [45, 46] for different values of relative submergence and for a limited range
of reduced velocity (U = 0.98 8).
In this range, it is possible to observe how the relative submergence influences
the dynamic response of the system, the frequency ratio f increases with h/D.

Figure 21: The influence of relative submergence h/D for vibrating spheres.

Flow Induced Excitation on Basic Shape Structures

21

This behavior can be shown through the effect that the deformation of the free
surface has on the oscillations of the sphere. In fact, for values of h/D = 1 the free
surface deforms and the vortex layer is generated between the free surface and
the upper obstacle surface. This layer gives rise to a near-wake conditioning the
frequency body response. Vortex generation selects frequency ranges, which can
include the proper obstacle frequency and can involve typical aspects related to
the locking-in effects. Values of h/D > 1, on the other hand, pull the system away
from the condition of lock-in and synchronization.

4 Conclusion
Flow induced excitations on structures represent a relevant and topic related to
several modern theoretical and practical engineering problems. The aim of this
chapter was to provide updated information about findings concerning two aspects
related to the excitation on vibrating structures. In fact, the approach followed
takes into account two main points of view on the processes: firstly, the flow field
and the effect due to the turbulent features of the wake have been discussed as
the source of the vibration on the structure. Secondly, the interaction between
flow and structure has been proposed in terms of dynamic response of the obstacle. The process of flow induced excitations focuses on the framework of causeeffect referring to basic-shape structures like circular, and/or rectangular cylinders
and spheres.

5 Acknowledgments
The authors gratefully acknowledge the financial support of the Italian Ministry of
Scientific and Technology Research, for the project PRIN 2002 entitled Influence
of vorticity and turbulence in interactions of water bodies with their boundary
elements and effect on hydraulic design.
Further, sincere thanks to the professors Naudascher & Rockwell for the use of
some of their figures, plots and information as well as the anonymous reviewers for
their kind and precious support during the writing of the chapter.
List of symbols
coefficient of mechanical damping [M T 1 ]

= torsional damping coefficient [M L2 T 1 ]

= spring constant coefficient [M T 2 ]

C
C

CA
CD
CL
C

=
=

added stiffness [M T 2 ]
potential added mass coefficient [/]

drag coefficients [/]

=
=

lift coefficient [/]

torsional spring constant [M L2 T 2 ]

22 Vorticity and Turbulence Effects in Fluid Structure Interaction

D
f
f
fn
f0
fs
F
F0
FD
F rs
G
h
h
h/D

= characteristic dimension of body [L]

= oscillating frequency of the body [T 1 ]
=
=

frequency ratio (f /fn ) [/]

natural frequency of body oscillator [T 1 ]

frequency of vortex shedding [T 1 ]

= forcing frequency [T 1 ]
= fluid force on an obstacle [M LT 2 ]
=
=

force amplitude [M LT 2 ]
unit per length drag force [M T 2 ]

=
=

Froude number referred to the cylinder thickness [/]

elevation of the obstacle by the wall [L]

depth of water [L]

=
=

dimensionless parameter of obstacle submersion [/]

relative submergence [/]

= mass moment of inertia [M L2 ]

= length of body along the flow direction [L]

m
m

mass of body [M ]

mass ratio [/]

ma
M

= added mass [M ]
= moment [M L2 T 2 ]

M0
Re

= moment amplitude [M L2 T 2 ]
= Reynolds number [/]

SG

Sh
t
Tu
u
rms

= Strouhal number [/]

= time [T ]
= free stream turbulent level [/]

U0
Uw
x, y
x0 , y0
y
y0

Skop-Griffin parameter [/]

standard deviation of U0 on inflow main direction [LT 1 ]

=
=

normalized velocity [/]

mean velocity of the incoming flow [LT 1 ]

= average velocity in the wake [LT 1 ]

= stream-wise, transverse displacement [L]
= vibration amplitude in x,y direction [L]
= transverse displacement [L]

= vibration amplitude in y direction [L]

= boundary layer thickness [L]

blockage ratio [/]

Flow Induced Excitation on Basic Shape Structures

= fluid dynamic viscosity [M L1 T 1 ]

= fluid kinematic viscosity [L2 T 1 ]
= fluid density [M L3 ]

23

damping ratio [/]

= torsional damping ratio [/]

= angular or torsional displacement [/]

= amplitude of torsional vibration [/]

= phase angle [/]

= logarithmic decrement of mechanical damping [/]

= circular frequency [T 1 ]

d
n

circular frequency of damped oscillator [T 1 ]

circular natural frequency [T 1 ]

circular forcing frequency [T 1 ]

circular natural frequency of torsional vibration [T 1 ]

References
[1] Sarpkaya, T., Vortex-induced oscillations, Journal of Applied Mechanics,
46, pp. 241258, 1979.
[2] Ramberg, S.E. & Griffin, O.M., Hydroelastic response of marine cables
and risers, In Hydrodynamics in Ocean Engineering, Norwegian Institute of
Technology, Trondheim, Norway, pp. 12231245, 1981.
[3] Bearman, P.W., Vortex shedding from oscillating bluff bodies, Annual
Review of Fluid Mechanics, 16, pp. 195222, 1984.
[4] Blevins, R.D., Flow-induced vibrations, New York: Van Nostrand Reinhold,
1990.
[5] Naudascher, E. & Rockwell, D., Flow induced vibration: an engineering
guide, Rotterdam, Balkema, 1993.
[6] Bearman, P.W., Graham, J.M.R. & Obasaju, E.D., A study of forces, circulation and vortex patterns around a circular cylinder in oscillating flow,
Journal of Fluid Mechanics, 196, pp. 467494, 1988.
[7] Keulegan, G.M. & Carpenter, L.H., Forces on cylinders and plates in an
oscillanting fluid, Journal of Research of the National Bureau of Standards,
60(5), pp. 423440, 1958.
[8] Blackburn, H.M. & Henderson, R.D., A study of two-dimensional flow past
an oscillating cylinder, Journal of Fluid Mechanics, 385, pp. 255286, 1999.
[9] Lin, J.C., Vorobieff, P. & Rockwell, D., 3-dimensional patterns of streamwise vorticity in the turbulent near-wake of a cylinder, Journal of Fluids and
Structures, 9, pp. 231234, 1995.
[10] Sheridan, J., Lin, J.C. & Rockwell, D., Flow past a cylinder close to a free
surface, Journal of Fluid Mechanics, 300, pp. 130, 1997.

24 Vorticity and Turbulence Effects in Fluid Structure Interaction

[11] Naudascher, E., AIRH Design Manual: Hydrodynamic forces. Rotterdam:
A.A. Balkema Publishers, 1991.
[12] Guilmineau, E. & Queutey, P., A numerical simulation of vortex shedding
from an oscillating circular cylinder, Journal of Fluids and Structures, 16(6),
pp. 773794, 2002.
[13] Lam, K.M. & Dai, G.Q., Formation of vortex street and vortex pair from
a circular cylinder oscillating in water, Experimental Thermal and Fluid
Science, 26, pp. 901915, 1998.
[14] Yan, B., Oscillatory flow beneath a free surface, Fluid Dynamic Research,
22, pp. 123, 1998.
[15] Munson, B.R.,Young, D.F. & Okiishi, T.H., Fundamentals of Fluid Mechanics (third edition), John Wiley & Sons, Inc., 1998.
[16] Yu, D. & Kareem, A., Parametric study of flow around rectangular prisms
using LES, Journal Wind Engineering and Industrial Aerodynamics, 78,
pp. 653662, 1998.
[17] Naudasher, E. & Wang, Y., Flow-induced vibrations of prismatic bodies and
grids of prisms, Journal of Fluids and Structures, 7, pp. 341373, 1993.
[18] Deniz, S. & Staubli, Th., Oscillating rectangular and octagonal profiles: interaction of leading- and trailing-edge vortex formation, Journal of Fluids and
Structures, 11(1), pp. 331, 1997.
[19] Bearman, P.W. & Trueman, D.M., An investigation of the flow around rectangular cylinder, The Aeronautical Quarterly, 23, pp. 229237, 1972.
[20] Shimada, K. & Ishiara, T., Application of modified k-e model to the prediction of aerodynamic characteristics of rectangular cross section cylinders,
Journal of Fluids and Structures, 16(4), pp. 465485, 2002.
[21] Okajima, A., Strouhal numbers of rectangular cylinders, Journal of Fluids
Mechanics, 123, pp. 379398, 1982.
[22] Haan, F.L., Kareem, A. & Szewczyk, A.A., The effects of turbulence on the
pressure distribution around a rectangular prism, Journal of Wind Engineering and Industrial Aerodynamics, 78, pp. 381392, 1998.
[23] Lin, J.C. & Melbourne, W.H., Turbulence effects on surface pressure of
rectangular cylinders, Wind and Structures, 2(4), pp. 253266, 1999.
[24] Noda, H. & Nakayama, A., Free-stream turbulence effects on the instantaneous pressure and forces on cylinders of rectangular cross section, Experiments in Fluids, 34, pp. 332344, 2003.
[25] Saathoff, P.J. & Melbourne, W.H., Effects of free-stream turbulence on surface pressure fluctuations in a separation bubble, Journal of Fluids Mechanics, 337, pp. 124, 1997.
[26] Nakamura, Y., Ohia, Y., Ozono, S. & Nakayama, R., Experimental and numerical analysis of vortex shedding from elongated rectangular cylinders at
low Reynolds numbers 2001000, Journal of Wind Engineering and Industrial Aerodynamics, 65, pp. 301308, 1996.
[27] Price, S.J., Sumner, D., Smith, J.G., Leong, K. & Paig Doussis, M.P., Flow
visualization around a circular cylinder near to a plane wall, Journal of
Fluids and Structures, 16(2), pp. 175191, 2002.

Flow Induced Excitation on Basic Shape Structures

25

[28] Angrilli, F., Bergamaschi, S. & Cossalter, V., Investigation of wall induced
modifications to vortex shedding from a circular cylinder, ASME Journal of
Fluids Engineering, 104, pp. 518522, 1982.
[29] Bearman, P.W. & Zdravkovich, M.M., Flow around a circular cylinder near
a plane boundary, Journal of Fluid Mechanics, 89, pp. 3347, 1978.
[30] Buresti, G. & Lanciotti, A., Vortex shedding from smooth and roughened
cylinders in cross-flow near a plane surface, The Aeronautical Quarterly,
30, pp. 305321, 1979.
[31] Taniguchi, S. & Miyakoshi, K., Fluctuating fluid forces acting on a circular cylinder and interference with a plane wall, Experiments in Fluids, 9,
pp. 197204, 1990.
[32] Lei, C., Cheng, L. & Kavanagh, K., Re-examination of the effect of a plane
boundary on force and vortex shedding of a circular cylinder, Journal of
Wind Engineering and Industrial Aerodynamics, 80, pp. 263286, 1999.
[33] Cigada, A., Malavasi, S. & Vanali, M., Experimental studies on the boundary
condition effects on the flow around a rectangular cylinder, Fluid Structure
Interaction 2003, Cadiz, Spain, 2426 June, 2003.
[34] Malavasi, S. & Guadagnini, A., Hydrodynamic loading on river bridges,
Journal Hydraulic Engeneering (ASCE), 129(11), November 2003, pp. 854
861, 2003.
[35] Denson, K.H., Steady-state drag, lift, and rolling-moment coefficients for
inundated inland bridges, Rep. No. MSHD-RD-82-077, reproduced by
National Technical Information Service, Springfield, Virg., pp. 123, 1982.
[36] Malavasi, S., Franzetti, S. & Blois, G., PIV Investigation of Flow Around
Submerged River Bridge, Proc. of River Flow 2004, Napoli (Italy), June
2325, 2004.
[37] Feng, C.C., The measurement of vortex-induced effects in flow past a stationary and oscillating circular and D-section cylinders, Masters Thesis,
Univ. of British Columbia, Vancouver, B.C., Canada, 1968.
[38] Griffin, O.M., Vortex-excited cross-flow vibrations of a single cylindrical
tube, ASME Journal of Pressure Vessel Technology, 102, pp. 158166, 1980.
[39] Blackburn, H., & Karniadakis, G.E., Two and Three dimensional simulations of vortex-induced vibration of a circular cylinder, In 3rd International
Offshore and Polar Engineering Conference, 3, pp. 715720, 1993.
[40] Skop, R.A. & Balasubramanian, S., A new twist on an old model for vortexexcited vibrations, Journal of Fluids and Structures, 11, pp. 395412, 1997.
[41] Khalak, A., & Williamson, C.H.K., Motion, forces and mode transitions
in vortex-induced vibrations at low mass-damping, Journal of Fluids and
Structures, 13, pp. 813851, 1999.
[42] Govardhan, R. & Williamson, C.H.K., Modes of vortex formation and frequency response of a freely vibrating cylinder, Journal of Fluid Mechanics,
420, pp. 85130, 2000.
[43] Sarpkaya, T., Hydrodynamics damping, flow-induced oscillations, and biharmonic response, ASME Journal of Offshore Mechanics and Artic Engineering, 117, pp. 232238, 1995.

26 Vorticity and Turbulence Effects in Fluid Structure Interaction

[44] Jauvtis, N., Govardhan, R. & Williamson, C.H.K., Multiple modes of vortexinduced vibration of a sphere, Journal of Fluids and Structures, 15, pp. 555
563, 2001.
[45] Mirauda, D. & Greco, M., Transverse vibrations of an sphere at high combined mass-damping parameter, Shallow Flows Jirka and Uijittewaal (eds)
Balkema Publisher, Taylor and Francis Group, London, ISBN 90 5809 700
5, pp. 111116, 2004.
[46] Mirauda, D. & Greco, M., Flow-induced vibration of an elastically sphere
at high combined mass-damping parameter, Journal of IASME Transactions
on Mechanical Engineering, 1, pp. 486491, 2004.